A framework for mapping political networks using the tweets from members of a national legislature: the case of the German Bundestag

Harald Meier

Contact: harald@smrfoundation.org

Twitter has become a popular communication tool for politicians around the world. In this study we present a framework for the analysis of Twitter networks that form among the elected representatives within a national parliament using the German Bundestag as a case study. A broad set of research questions can be addressed with this approach: How does party affiliation align with network clustering? Which parties are most central, which are most peripheral? Which politicians are most influential? How do vertex centrality metrics align with party influence? Which Twitter accounts outside the Parliament are important to the overall network and to each party? Which networking strategies are pursued by each party? Which network metrics can be considered to measure intra-party cohesion? Over the past year we have collected monthly datasets with tweets by all Members of the German Bundestag via the public Twitter API and used the network analysis tool NodeXL Pro which performs social network analysis in combination with content analysis. Each dataset offers multiple layers of analysis: The full network view contains all network connections that were collected, while the internal network view only considers connections between the politicians themselves. Both of these perspectives are also analyzed on the party level. We find that internal party cohesion can be measured and varies significantly with Bündnis90/Die Grünen most internally cohesive and CDU/CSU least internally interconnected. The inter-party network shows the AfD is least connected and most peripheral within the Bundestag, while content analysis shows that #AfD is one of the most frequently used hashtag across party lines. The clustering algorithm (Clauset-Newman-Moore) detects different sets of party coalitions when considering different layers of network data and also different time frames: Black-Red (CDU/CSU and SPD), Black-Yellow (CDU/CSU and FDP), Red-Red (SPD and Die Linke), Green-Red (Bündnis90/Die Grünen and Die Linke). Centrality metrics reveal that top party representatives and politicians with a high tweet frequency also occupy the most influential and central locations in the network. The most important Twitter actors outside the Bundestag are large media outlets, party handles, as well as regional and national politicians without Bundestag mandates. In the next step, we will expand our study to other Parliaments around the world to be able to compare the findings from this study to establish baseline datasets for the public social media behavior of democratic institutions.

← Schedule